Lexical Semantics and Knowledge Representation


In Berkeley, California, USA with ACL 1991

Endorsed by SIGLEX.

The recent resurgence of interest in lexical semantics (LS) has brought many linguistic formalisms closer to the knowledge representation (KR) languages utilized in AI. In fact, some formalisms from computational linguistics are emerging which may be more expressive and formally better understood than many KR languages. Furthermore, the interests of computational linguists now extend to include areas previously thought beyond the scope of grammar and linguistics, such as commonsense knowledge, inheritance, default reasoning, collocational relations, and even domain knowledge.

With such an extension of the purview of “linguistic” knowledge, the question emerges as to whether there is any logical justification for distinguishing between lexical semantics and world knowledge. The purpose of this workshop is to explore this question in detail, with papers addressing the following points:

  1. Possible methods for determining what is lexical knowledge and what is outside the scope of such knowledge.
  2. Potential demonstrations that the inferences necessary for language understanding are no different from supposed non-linguistic inferences.
  3. Arguments from language acquisition and general concept development.
  4. Cross-linguistic evidence for the specificity of lexical semantic representations.
  5. Philosophical arguments for the (impossibility of the) autonomy of lexical knowledge.
  6. Theoretical approaches and implemented systems that combine lexical and non-lexical knowledge.